Human papillomavirus genotype and cycle threshold value from self-samples and risk of high-grade cervical lesions: A post hoc analysis of a modified stepped-wedge implementation feasibility trial.

Background: Human papillomavirus (HPV) testing of self-collected vaginal samples has potential to improve coverage of cervical screening programmes, but current guidelines mostly require those HPV positive on a self-sample to attend for routine screening.

Methods and findings: A pragmatic modified stepped-wedge implementation feasibility trial was conducted at primary care practices in England. Individuals aged 25 to 64 years who were at least 6 months overdue for cervical screening could provide a self-collected sample. The primary outcomes included the monthly proportion of non-attenders screened, changes in coverage, and uptake within 90 days. Self-samples from 7,739 individuals were analysed using Roche Cobas 4800. Individuals with a positive self-sample were encouraged to attend clinical screening. In this post hoc study of the trial, we related the HPV type (HPV16, HPV18, or other high-risk type) and cycle threshold (Ct) value on the self-sample to the results of clinician-collected sample and cervical intraepithelial neoplasia grade 2 or worse (CIN2+). We wished to triage HPV-positive individuals to immediate colposcopy, clinician sampling, or 12-month recall depending on risk. A total of 1,001 women tested positive through self-samples, and 855 women who had both an HPV-positive self-sample and a subsequent clinician-sample were included in this study. Of these, 71 (8.3%) had CIN2+. Self-sample Ct values were highly predictive of HPV in the clinician sample. Combining HPV type and Ct value allowed stratification into 3 risk groups; 44/855 (5%) were high-risk of whom 43% (19/44, 95% confidence interval [29.7%, 57.8%]) had CIN2+. The majority (52.9%, 452/855) were low-risk, of whom 4% (18/452, 95% CI [2.5%, 6.2%]) had CIN2+. The main limitation of our study was the colposcopy assessment was restricted to individuals who had abnormal cytology after positive results of both self-sample and clinician-collected sample.

Conclusions: HPV type and Ct value on HPV-positive self-samples may be used for triage. The difference in the risk of CIN2+ in these groups appears sufficient to justify differential clinical management. A prospective study employing such triage to evaluate laboratory workflow, acceptability, and follow-up procedure and to optimise clinical performance seems warranted.

Trial registration: ISRCTN12759467.